---
license: apache-2.0
base_model:
- facebook/detr-resnet-50
pipeline_tag: object-detection
tags:
- object-detection
- computer-vision
- renesas
- x5h
- onnx
- detr
- transformer
- detection
---

# DETR-R50 (ONNX) – Renesas X5H

## Introduction

This repository hosts **DETR (DEtection TRansformer)**, targeting the **Renesas R-Car X5H**
platform for object detection inference on the NPX6 NPU.

- **Model Architecture:** DETR — transformer-based, end-to-end object detector using set
  prediction (bipartite matching), CNN backbone + transformer encoder/decoder
- **Source Model:** [facebook/detr-resnet-50](https://huggingface.co/facebook/detr-resnet-50) — checkpoint [`sim_mod_detr`](https://huggingface.co/facebook/detr-resnet-50)
- **Task:** Object Detection
- **Dataset:** COCO (inferred from checkpoint name)
- **Input Resolution:** TBD

## Deployment Flow

The FP32 ONNX model is auto-cast to **INT8** by the Renesas MWMX toolchain at compile time — no
separate quantization step is required.

```
sim_mod_detr_..._optimized.onnx (FP32)
        │
        └─▶  MWMX Runtime  ──▶  INT8 auto-cast  ──▶  NPX6 NPU
```

## Provided Artifacts

| Artifact | Status | Notes |
|----------|--------|-------|
| **FP32 (ONNX)** | ✅ | `fp32/sim_mod_detr.onnx` — FP32 ONNX export |

## Performance

Measured on **Renesas R-Car X5H** via the MWMX runtime (APM50 ship-performance CI pipeline).

> **Benchmark configuration:** Single NPU · Batch size: 1 · Input resolution: TBD

| Runtime | Precision | Device | Latency (ms) | Type |
|---------|-----------|--------|---------------|------|
| MWMX Runtime | INT8 (auto) | X5H · 1× NPU · 1 Core · 850 MHz | 147.768855 | Measured |
| MWMX Runtime | INT8 (auto) | X5H · 1× NPU · 12 Cores · 850 MHz | 92.763967 | Measured |

---

## Model Input

### Input Tensor

- Shape: TBD — not available from source data (expected `(N, 3, H, W)`, RGB)
- Format: TBD
- Data Type: TBD
- Pixel Range: TBD

### Preprocessing

TBD — not available from source data.

## Model Outputs

TBD — not available from source data. DETR produces a fixed-size set of predictions (typically
100 object queries), each with a class-probability distribution (including a "no object" class)
and a normalized bounding box, directly via set prediction — no anchor decoding or NMS is
required by design.

### Postprocessing

1. Per-query class-probability argmax (excluding "no object")
2. Confidence threshold filtering
3. Box de-normalization to image coordinates

### Accuracy

TBD — not yet measured/published for this repo.

---

## Runtime Details

### MWMX Runtime

- **Engine:** Renesas MWMX (Middleware MX) native inference runtime
- **Input format:** FP32 ONNX (compiled by the MWMX toolchain)
- **NPU execution precision:** INT8 (auto-cast by MWMX toolchain)
- **Execution target:** NPX6-48K NPU on R-Car X5H

---

## Prerequisites

To run inference on Renesas R-Car X5H, you need:

1. **Renesas R-Car X5H board** with NPX6 NPU
2. **Renesas MWMX Runtime**
3. **Hugging Face CLI** to download the model

## Download

```bash
hf download Renesas/DETR-R50-ONNX --repo-type=model --include "fp32/*"
```

---

## Benchmark Methodology

- **HIL runs:** Hardware-in-the-loop — measured on physical R-Car X5H silicon via the MWMX
  runtime (`metawaremx_runtime` CI pipeline, "APM50" ship-performance target)
- **Precision:** FP32 ONNX input; INT8 execution (auto-cast by MWMX)
- **Slices:** results reported for both 1 AI core and 12 AI cores per NPU instance